Kevin, I did one step more. I did a ceiling bounce with a white door being photographed in manual mode. I used a p60 with new batteries for reference, the l2, u2 and u2 + l2.
I did a 25x25 point sample at a point in each picture to derive the rgb. I summed the components and applied a ratio to the known approximate output results.
P60=1*65lumens=65 lumens (hard math) (22 was the r+g+b score)
L2 rgb score =33 33/22 * 65 = 97.5 lumens
U2 rgb score =41 41/22 * 65 = 121 lumens
u2+l2 rgb score =78 78/22 * 65 = 230 lumens
Of course there is error in the ccd due to the bayer filter and other issues. The system isn't technically sound but is isn't a bad estimation assuming the p60=65lumens was a good calibration. So the simple test does help to back my claim of the difference between the two.
I'd imagine I can make it work better, including giving spectral results but it was more of a verification. I need to do the same without he B65 battery and also the 3x123 u2 which seems to give about a 20lumen increase (just a test not a long term burn because I like the u2 and I only have 3 levels vs 6 when doing this)